Mechanisms of ATP-Dependent Chromatin Remodeling Motors

Abstract

Chromatin remodeling motors play essential roles in all DNA-based processes. These motors catalyze diverse outcomes ranging from sliding the smallest units of chromatin, known as nucleosomes, to completely disassembling chromatin. The broad range of actions carried out by these motors on the complex template presented by chromatin raises many stimulating mechanistic questions. Other well-studied nucleic acid motors provide examples of the depth of mechanistic understanding that is achievable from detailed biophysical studies. We use these studies as a guiding framework to discuss the current state of knowledge of chromatin remodeling mechanisms and highlight exciting open questions that would continue to benefit from biophysical analyses.

Keywords: ATPase; INO80; ISWI; SWI/SNF; SWR; histones; molecular motors; nucleosome.

Figure 1

Case studies of nucleic acid motors. (a) Crystal structure of a RecA filament bound to single-stranded DNA (red ribbon) with ADP-AlF₄ (blue space-fill) [Protein Data Bank ID (PDB ID): 3CMU] (18). (b) Domain architecture of SF1 and SF2 ATPases. The conserved RecA-like folds (orange and tan) can be elaborated by insertions or modifications between the RecA folds (gray triangle) or within either RecA-like fold (blue triangles), as well as by N- and C-terminal extensions (green and yellow, respectively). In panels c and d, for simplicity, schematics show only N- and C-terminal extensions without indication of insertions and modifications in the RecA folds. (c) Structure of Mss116 bound to duplex RNA and AMP-PNP (PDB ID: 3i5x) (32). The N-terminal extension is disordered in the crystal structure and is not shown. (d) Structure of the RecBCD complex and partially unwound DNA (red ribbons) with the pin motif in magenta (PDB ID: 3K70) (112). (e) Crystal structure of the N-terminal domain of Modifier of transcription 1 (Mot1) bound to TATA-binding protein (TBP), with the approximate location of the ATPase domain, based on electron microscopy and cross-linking data, in gray (PDB ID: 3OC3) (141). Although Mot1’s mechanism is not fully understood, movement of the RecA-fold-containing lobes of the ATPase is likely coupled to TBP removal. Crystal structure representations in all figures were made with UCSF Chimera (http://www.cgl.ucsf.edu/chimera). Abbreviations: dsRNA, double-stranded RNA; Pi, inorganic phosphate; ssRNA, single-stranded RNA.

Figure 2

General properties of nucleosomes. (a) Two space-filling views of a nucleosome core particle (PDB ID: 1KX5), with the DNA in gray and the histone proteins in red/orange (28). (b) Schematic of all physiologically accessible states for a nucleosome observed to date. H2A/H2B dimers are shown in orange and the H3/H4 tetramer is shown in red. (c) Top view of a nucleosome with histones and superhelical locations (SHL) on the DNA labeled. SHL ± 2, the binding location for several remodeler families, is highlighted in yellow. (d) Hypothetical free-energy profiles for a nucleosome remodeling reaction. The same remodeled product can be made through a physiologically accessible (right) or alternate (left) path.

Figure 3

Model for ISWI remodeling. (a) Domain architecture of ISWI-family ATPase subunits, with the names of ISWI ATPases and complexes that are referenced in this review. (b) A model for directional nucleosome sliding by the human ISWI motor SNF2h (adapted from Reference with permission). The nucleosome is represented from a top-down perspective, and for simplicity, only one protomer of SNF2h is shown. The ATPase domain of SNF2h binds at SHL ± 2 and engages the H4 tail. Binding of the HSS domain to flanking DNA (HSS-out) stimulates ATP hydrolysis, driving SNF2h into a translocation competent state (HSS-in). Translocation (in the direction of red arrows) may be coupled to distortion of the nucleosome. Abbreviations: HSS, HAND-SANT-SLIDE; Pi, inorganic phosphate. (c) Model for additional regulation by Acf1 in the pause phase (58). When flanking DNA is short, Acf1 sequesters the H4 tail, inhibiting exit from the pause. When flanking DNA is long, the N-terminus of Acf1 instead binds to flanking DNA, allowing exit from the pause.

Figure 4

Model for SWR remodeling. (a) Model for directional histone exchange by SWR (adapted from Reference with permission). The Swr1 ATPase (brown) engages a nucleosome containing a canonical H2A/H2B dimer (orange) at SHL ± 2 on the distal side relative to flanking DNA (107). The Swc2 subunit and SWR1 both engage an H2A.Z/H2B dimer (blue). When both the nucleosome and H2A.Z are bound, Swr1’s ATPase activity is maximally stimulated and a single H2A/H2B dimer is replaced by an H2A.Z/H2B dimer. Swr1 may also act processively until both canonical dimers have been replaced by H2A.Z/H2B dimers. (b) Domain architectures of the Ino80 and Swr1 ATPases. (c) The INO80 and SWR complexes may share a common intermediate in which DNA is unwrapped from one side of the nucleosome. SWR may use this intermediate to facilitate histone exchange with H2A.Z/H2B, whereas INO80 may use it to slide the histone octamer toward the site of unwrapping. Abbreviations: HSA, helicase-SANT–associated; Pi, inorganic phosphate.

Figure 5

Model for SWI/SNF remodeling. (a) Domain architecture of SWI/SNF-family ATPase subunits, with the names of SWI/SNF ATPases and complexes that are referenced in this review. (b) The SWI/SNF ATPase domain engages the nucleosome at SHL ± 2 (110, 152). ATP-dependent translocation by SWI/SNF disrupts histone-DNA contacts across the nucleosome. From this disrupted intermediate, the nucleosome may collapse into several stable products, including (from left to right) repositioned nucleosomes, nucleosomes with stable surface loops, nucleosomes with exchanged dimers, and octamers transferred to free DNA. SWI/SNF can also generate a dinucleosome-like species (far right) between two nucleosomes formed on separate DNA templates, in which one DNA strand bridges two octamers (76, 80, 101, 113). A key aspect of this model is that most or all of these products can be recycled back to the disrupted intermediate through the action of the motor. Abbreviations: HSA, helicase-SANT–associated; Pi, inorganic phosphate.

Figure 6

In vivo, different families of remodeling enzymes collaborate with one another and with other nuclear factors to regulate gene expression. The mechanisms behind this collaboration are not well understood and are an area of continuing interest in the field. Abbreviation: Pi, inorganic phosphate.